The proposed research plan applies recent theoretical models to the variety of cognitive changes that characterize Dementia of the Alzheimer's Type (DAT). Healthy elderly adults exhibit changes that appear similar to, but much milder than, those observed in DAT. In three studies, each examining performance on at least 12 tasks, subjects with DAT will be compared with both young and healthy elderly adults in order to address the question of whether cognitive changes in DAT represent exaggerated aging or unique pathology. Taken together, the three studies will generate precise information about the interrelations between many of the cognitive symptoms of this disease, focusing on the extent to which slowing in DAT is global or task-specific, and the consequences of slower processing for memory and cognition. Salthouse has proposed a model in which the cognitive changes associated with normal aging form a cascade: A global decrease in processing speed leads to a decrease in working memory capacity that, in turn, causes a decline in higher cognitive abilities. The purposes of the first study are to test a model of cognitive changes in DAT that parallels Salthouse's account of normal aging and to explore the relations between speed, memory, and higher cognitive abilities. The second study will examine more closely the link between speed and memory. Baddeley has proposed a three-component model in which working memory function depends upon two modality-specific slave systems controlled by a central executive. Like Salthouse, Baddeley has suggested that speed, more specifically articulation rate, is a major determinant of the capacity of working memory. In addition, Baddeley suggested that deficits in the central executive underlie working memory deficits in DAT. The second study will evaluate Baddeley's model and test his hypothesis concerning central executive deficits in DAT. In healthy elderly, the ability to process and remember lexical information is much better preserved than the ability to process and remember nonlexical information, and the same appears to be true in DAT. Although Baddeley's working memory model includes only one slave system for visuospatial information, recent research in cognitive neuroscience has identified separate neural processing streams specialized for perceiving and remembering objects and locations. The third study will examine whether object and location information are both more sensitive than lexical information to the effects of aging and DAT.